Re: forbidden, forbidden

At 14:12 18.05.98 -0400, Richard L. Goerwitz III wrote:
>Another perhaps silly question:
>
>Why are so many things explicitly forbidden in the table in section 4.4
>of the XML spec (the table that summarizes where what sorts of entities
>can occur, and what should happen to them)?
>
>I'm just thinking about this as a programmer.
>
>Why not just ignore some of these things, if they occur in a context
>where they have no meaning or function.  Later, if you change your minds
>about their meaning and function, you can change the spec.  And docs
>that were once valid will remain valid.  Docs that were in error will
>still be in error.

Well, no.  If we changed all the 'forbidden' items to 'legal but
ignored', and later changed our minds about the meaning and function 
of some entity usage, then well-formed documents might still be
well-formed, but they would suddenly have a different interpretation.
I.e. adding a category of 'ignored' does not provide a graceful 
migration path, after all.  In fact, I think the migration from
'forbidden' to a defined use and function is a more graceful path:
existing well-formed or valid documents are not affected at all,
and previously ill-formed documents gain an interpretation as
well-formed documents.

Just my opinion.

-C. M. Sperberg-McQueen

Received on Tuesday, 19 May 1998 02:51:49 UTC